Objective: The aim of this study was to assess the extent of misreporting in obese and nonobese adults on an absolute, ratio-scaled, and allometrically-scaled basis.
Method: Self-reported daily energy intake (EI) was compared with total energy expenditure (TEE) in 221 adults (106 male, 115 female; age 53 ± 17 years, stature 1.68 ± 0.09 m, mass 79.8 ± 17.2 kg) who participated in a doubly-labeled water (DLW) subsection of 2013-2015 National Diet and Nutrition Survey. Data were log transformed and expressed as absolute values, according to simple ratio-standards (per kg body mass) and adjusted for body mass allometrically. Absolute and ratio-scaled misreporting were examined using full-factorial General Linear Models with repeated measures of the natural logarithms of TEE or EI as the within-subjects factor. The natural logarithm of body mass was included as a covariate in the allometric method. The categorical variables of gender, age, obesity, and physical activity level (PAL) were the between-factor variables.
Results: On an absolute-basis, self-reported EI (2759 ± 590 kcal·d-1 ) was significantly lower than TEE measured by DLW (2759 ± 590 kcal·d-1 : F1,205 = 598.81, p < .001, ηp 2 =0.75). We identified significantly greater underreporting in individuals with an obese BMI (F1,205 = 29.01, p <.001, ηp 2 =0.12), in more active individuals (PAL > 1.75; F1,205 = 34.15, p <.001, ηp 2 =0.14) and in younger individuals (≤55 years; F1,205 = 14.82, p < .001, ηp 2 =0.07), which are all categories with higher energy needs. Ratio-scaling data reduced the effect sizes. Allometric-scaling removed the effect of body mass (F1,205 =0.02, p = 0.887, ηp 2 =0.00).
Conclusion: In weight-stable adults, obese individuals do not underreport dietary intake to a greater extent than nonobese individuals. These results contradict previous research demonstrating that obesity is associated with a greater degree of underreporting.
© 2022 The Authors. American Journal of Human Biology published by Wiley Periodicals LLC.